httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ed Cooley" <ecoo...@strategy5.com>
Subject Re: [users@httpd] apache signals
Date Tue, 07 Dec 2004 21:51:33 GMT
Thanks for the feedback.

 

What is supposed to be sending SIGPIPE signals?  Is apache doing that or is
the OS?  We don't seem to be getting any SIGPIPE signals.

 

I have a test CGI that outputs for 30 seconds and while it is processing we
request the same cgi from the same browser window (hitting F5).  Now I have
two instances executing the second outputs and displays, the first one
finishes and gets no SIGPIPE and of course the browser never displays it's
output.  We are thinking the first executable should get some kind of signal
so it can be shut down.

 

Ed.

 

-----Original Message-----

From: Joshua Slive [mailto:jslive@gmail.com]

Sent: Tuesday, December 07, 2004 11:45 AM

To: users@httpd.apache.org

Subject: Re: [users@httpd] apache signals

 

On Tue, 7 Dec 2004 11:41:35 -0600, Ed Cooley <ecooley@strategy5.com> wrote:

> Where I can find documentation on how apache terminates a cgi that is 

> superseded by a duplicate request from the same browser window.  I 

> have searched for signal handling and cannot find anything.

 

"Request from the same browser window" means practically nothing to apache.
It can't tell and doesn't care which browser window a request comes from.

 

If the CGI tries to write content back to the client and the client is no
longer there, a SIGPIPE should result, which should kill the CGI. 

Otherwise, there won't be any termination until the CGI script is finished
or the regular timeouts elapse.

 

Joshua.

 

 


Mime
View raw message